DESCRIPTION: (Applicant's Description) Numerous smoking cessation treatments have been developed over the past 20 years. Unfortunately, these have not led to a notable increase in cessation rates over the same period. One impediment to a rational, empirically-based approach to smoking treatment is that we know very little about how effective treatments work. The proposed research will use Electronic Diaries (EDs; palmtop computers) to gather fine-grained, virtually real-time information on variables thought to mediate the effects of efficacious smoking cessation treatments. In addition, recent research suggests that tonic change in affect/withdrawal powerfully predicts relapse; this project will permit sensitive tests of this relation. In this research, 390 smokers motivated to quit will be randomized to one of three groups. In the Bupropion group, smokers will attempt to quit smoking with the aid of bupropion pharmacotherapy (150 mg b.i.d.). In the Combined condition, smokers will receive bupropion pharmacotherapy and will participate in group therapy sessions comprising therapeutic elements of known efficacy (i.e., social support and coping-skills training). A final group of smokers will be assigned to a Placebo condition; these subjects will not attend group sessions and will be given placebo pharmacotherapy. All subjects will use EDs to collect real-time information about urges, withdrawal, negative affect, smoking rates, etc -- during a 4-day period prior to the quit date and for the first 8 weeks of the cessation attempt. The study design allows for sensitive descriptive and predictive examinations of both short-term episodes and longer-term trends in theoretically prominent variables such as negative affect, withdrawal, stressors, temptations, and urges. It also permits tests of mediation to determine whether bupropion pharmacotherapy and/or counseling work through modifying the timing, frequency, trends, and/or severity of such variables.